idnits 2.17.1 draft-ietf-precis-problem-statement-08.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (September 19, 2012) is 4209 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-09) exists of draft-iab-identifier-comparison-04 -- Obsolete informational reference (is this intentional?): RFC 3454 (Obsoleted by RFC 7564) -- Obsolete informational reference (is this intentional?): RFC 3490 (Obsoleted by RFC 5890, RFC 5891) -- Obsolete informational reference (is this intentional?): RFC 3491 (Obsoleted by RFC 5891) -- Obsolete informational reference (is this intentional?): RFC 3530 (Obsoleted by RFC 7530) -- Obsolete informational reference (is this intentional?): RFC 3920 (Obsoleted by RFC 6120) -- Obsolete informational reference (is this intentional?): RFC 4013 (Obsoleted by RFC 7613) -- Obsolete informational reference (is this intentional?): RFC 5661 (Obsoleted by RFC 8881) Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 8 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group M. Blanchet 3 Internet-Draft Viagenie 4 Intended status: Informational A. Sullivan 5 Expires: March 23, 2013 Dyn, Inc. 6 September 19, 2012 8 Stringprep Revision and PRECIS Problem Statement 9 draft-ietf-precis-problem-statement-08.txt 11 Abstract 13 If a protocol expects to compare two strings and is prepared only for 14 those strings to be ASCII, then using Unicode codepoints in those 15 strings requires they be prepared somehow. Internationalizing Domain 16 Names in Applications (here called IDNA2003) defined and used 17 Stringprep and Nameprep. Other protocols subsequently defined 18 Stringprep profiles. A new approach different from Stringprep and 19 Nameprep is used for a revision of IDNA2003 (called IDNA2008). Other 20 Stringprep profiles need to be similarly updated or a replacement of 21 Stringprep needs to be designed. This document outlines the issues 22 to be faced by those designing a Stringprep replacement. 24 Status of this Memo 26 This Internet-Draft is submitted in full conformance with the 27 provisions of BCP 78 and BCP 79. 29 Internet-Drafts are working documents of the Internet Engineering 30 Task Force (IETF). Note that other groups may also distribute 31 working documents as Internet-Drafts. The list of current Internet- 32 Drafts is at http://datatracker.ietf.org/drafts/current/. 34 Internet-Drafts are draft documents valid for a maximum of six months 35 and may be updated, replaced, or obsoleted by other documents at any 36 time. It is inappropriate to use Internet-Drafts as reference 37 material or to cite them other than as "work in progress." 39 This Internet-Draft will expire on March 23, 2013. 41 Copyright Notice 43 Copyright (c) 2012 IETF Trust and the persons identified as the 44 document authors. All rights reserved. 46 This document is subject to BCP 78 and the IETF Trust's Legal 47 Provisions Relating to IETF Documents 48 (http://trustee.ietf.org/license-info) in effect on the date of 49 publication of this document. Please review these documents 50 carefully, as they describe your rights and restrictions with respect 51 to this document. Code Components extracted from this document must 52 include Simplified BSD License text as described in Section 4.e of 53 the Trust Legal Provisions and are provided without warranty as 54 described in the Simplified BSD License. 56 Table of Contents 58 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5 59 2. Keywords . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 60 3. Conventions . . . . . . . . . . . . . . . . . . . . . . . . . 6 61 4. Stringprep Profiles Limitations . . . . . . . . . . . . . . . 6 62 5. Major Topics for Consideration . . . . . . . . . . . . . . . . 8 63 5.1. Comparison . . . . . . . . . . . . . . . . . . . . . . . . 8 64 5.1.1. Types of Identifiers . . . . . . . . . . . . . . . . . 8 65 5.1.2. Effect of comparison . . . . . . . . . . . . . . . . . 8 66 5.2. Dealing with characters . . . . . . . . . . . . . . . . . 9 67 5.2.1. Case folding, case sensitivity, and case 68 preservation . . . . . . . . . . . . . . . . . . . . . 9 69 5.2.2. Stringprep and NFKC . . . . . . . . . . . . . . . . . 9 70 5.2.3. Character mapping . . . . . . . . . . . . . . . . . . 10 71 5.2.4. Prohibited characters . . . . . . . . . . . . . . . . 10 72 5.2.5. Internal structure, delimiters, and special 73 characters . . . . . . . . . . . . . . . . . . . . . . 10 74 5.2.6. Restrictions because of glyph similarity . . . . . . . 11 75 5.3. Where the data comes from and where it goes . . . . . . . 11 76 5.3.1. User input and the source of protocol elements . . . . 11 77 5.3.2. User output . . . . . . . . . . . . . . . . . . . . . 11 78 5.3.3. Operations . . . . . . . . . . . . . . . . . . . . . . 12 79 6. Considerations for Stringprep replacement . . . . . . . . . . 13 80 7. Security Considerations . . . . . . . . . . . . . . . . . . . 14 81 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 14 82 9. Discussion home for this draft . . . . . . . . . . . . . . . . 14 83 10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 14 84 11. Informative References . . . . . . . . . . . . . . . . . . . . 14 85 Appendix A. Classification of Stringprep Profiles . . . . . . . . 18 86 Appendix B. Evaluation of Stringprep Profiles . . . . . . . . . . 19 87 B.1. iSCSI Stringprep Profile: RFC3722 (and RFC3721, 88 RFC3720) . . . . . . . . . . . . . . . . . . . . . . . . . 19 89 B.2. SMTP/POP3/ManageSieve Stringprep Profiles: 90 RFC4954,RFC5034,RFC 5804 . . . . . . . . . . . . . . . . . 21 91 B.3. IMAP Stringprep Profiles: RFC5738, RFC4314: Usernames . . 22 92 B.4. IMAP Stringprep Profiles: RFC5738: Passwords . . . . . . . 24 93 B.5. Anonymous SASL Stringprep Profiles: RFC4505 . . . . . . . 25 94 B.6. XMPP Stringprep Profiles: RFC3920 Nodeprep . . . . . . . . 27 95 B.7. XMPP Stringprep Profiles: RFC3920 Resourceprep . . . . . . 28 96 B.8. EAP Stringprep Profiles: RFC3748 . . . . . . . . . . . . . 28 97 Appendix C. Changes between versions . . . . . . . . . . . . . . 29 98 C.1. 00 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 99 C.2. 01 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 100 C.3. 02 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 101 C.4. 03 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 102 C.5. 04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 103 C.6. 05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 104 C.7. 06 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 105 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 30 107 1. Introduction 109 Internationalizing Domain Names in Applications (here called 110 IDNA2003) [RFC3490], [RFC3491], [RFC3492], [RFC3454] describes a 111 mechanism for encoding Unicode labels making up Internationalized 112 Domain Names (IDNs) as standard DNS labels. The labels were 113 processed using a method called Nameprep [RFC3491] and Punycode 114 [RFC3492]. That method was specific to IDNA2003, but is generalized 115 as Stringprep [RFC3454]. The general mechanism is used by other 116 protocols with similar needs, but with different constraints than 117 IDNA2003. 119 Stringprep defines a framework within which protocols define their 120 Stringprep profiles. Known IETF specifications using Stringprep are 121 listed below: 122 o The Nameprep profile [RFC3490] for use in Internationalized Domain 123 Names (IDNs); 124 o NFSv4 [RFC3530] and NFSv4.1 [RFC5661]; 125 o The iSCSI profile [RFC3722] for use in Internet Small Computer 126 Systems Interface (iSCSI) Names; 127 o EAP [RFC3748]; 128 o The Nodeprep and Resourceprep profiles [RFC3920] for use in the 129 Extensible Messaging and Presence Protocol (XMPP), and the XMPP to 130 CPIM mapping [RFC3922] (the latter of these relies on the former); 131 o The Policy MIB profile [RFC4011] for use in the Simple Network 132 Management Protocol (SNMP); 133 o The SASLprep profile [RFC4013] for use in the Simple 134 Authentication and Security Layer (SASL), and SASL itself 135 [RFC4422]; 136 o TLS [RFC4279]; 137 o IMAP4 using SASLprep [RFC4314]; 138 o The trace profile [RFC4505] for use with the SASL ANONYMOUS 139 mechanism; 140 o The LDAP profile [RFC4518] for use with LDAP [RFC4511] and its 141 authentication methods [RFC4513]; 142 o Plain SASL using SASLprep [RFC4616]; 143 o NNTP using SASLprep [RFC4643]; 144 o PKIX subject identification using LDAPprep [RFC4683]; 145 o Internet Application Protocol Collation Registry [RFC4790]; 146 o SMTP Auth using SASLprep [RFC4954]; 147 o POP3 Auth using SASLprep [RFC5034]; 148 o TLS SRP using SASLprep [RFC5054]; 149 o IRI and URI in XMPP [RFC5122]; 150 o PKIX CRL using LDAPprep [RFC5280]; 151 o IAX using Nameprep [RFC5456]; 152 o SASL SCRAM using SASLprep [RFC5802]; 153 o Remote management of Sieve using SASLprep [RFC5804]; 154 o The unicode-casemap Unicode Collation [RFC5051]. 156 However, a review (see [ietf78precis]) of these protocol 157 specifications found that they are very similar and can be grouped 158 into a short number of classes. Moreover, many reuse the same 159 Stringprep profile, such as the SASL one. 161 IDNA2003 was replaced because of some limitations described in 162 [RFC4690]. The new IDN specification, called IDNA2008 [RFC5890], 163 [RFC5891], [RFC5892], [RFC5893] was designed based on the 164 considerations found in [RFC5894]. One of the effects of IDNA2008 is 165 that Nameprep and Stringprep are not used at all. Instead, an 166 algorithm based on Unicode properties of codepoints is defined. That 167 algorithm generates a stable and complete table of the supported 168 Unicode codepoints for each Unicode version. This algorithm is based 169 on an inclusion-based approach, instead of the exclusion-based 170 approach of Stringprep/Nameprep. That is, IDNA2003 created an 171 explicit list of excluded or mapped-away characters; anything in 172 Unicode 3.2 that was not so listed could be assumed to be allowed 173 under the protocol. IDNA2008 begins instead from the assumption that 174 code points are disallowed, and then relies on Unicode properties to 175 derive whether a given code point actually is allowed in the 176 protocol. 178 This document lists the shortcomings and issues found by protocols 179 listed above that defined Stringprep profiles. It also lists the 180 requirements for any potential replacement of Stringprep. 182 2. Keywords 184 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 185 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 186 document are to be interpreted as described in [RFC2119]. 188 3. Conventions 190 A single Unicode code point in this memo is denoted by "U+" followed 191 by four to six hexadecimal digits, as used in [Unicode61], Appendix 192 A. 194 4. Stringprep Profiles Limitations 196 During IETF 77 (March 2010), a BOF discussed the current state of the 197 protocols that have defined Stringprep profiles [NEWPREP]. The main 198 conclusions from that discussion were as follows: 199 o Stringprep is bound to version 3.2 of Unicode. Stringprep has not 200 been updated to new versions of Unicode. Therefore, the protocols 201 using Stringprep are stuck at Unicode 3.2, and their 202 specifications need to be updated to support new versions of 203 Unicode. 204 o The protocols would like to not be bound to a specific version of 205 Unicode, but rather have better Unicode version agility in the way 206 of IDNA2008. This is important partly because it is usually 207 impossible for an application to require Unicode 3.2; the 208 application gets whatever version of Unicode is available on the 209 host. 210 o The protocols require better bidirectional support (bidi) than 211 currently offered by Stringprep. 212 o If the protocols are updated to use a new version of Stringprep or 213 another framework, then backward compatibility is an important 214 requirement. For example, Stringprep is based on and profiles may 215 use NFKC [UAX15], while IDNA2008 mostly uses NFC [UAX15]. 216 o Identifiers are passed between protocols. For example, the same 217 username string of codepoints may be passed between SASL, XMPP, 218 LDAP and EAP. Therefore, common set of rules or classes of 219 strings are preferred over specific rules for each protocol. 220 Without real planning in advance, many stringprep profiles reuse 221 other profiles, so this goal was accomplished by accident with 222 Stringprep. 224 Protocols that use Stringprep profiles use strings for different 225 purposes: 226 o XMPP uses a different Stringprep profile for each part of the XMPP 227 address (JID): a localpart which is similar to a username and used 228 for authentication, a domainpart which is a domain name and a 229 resource part which is less restrictive than the localpart. 230 o iSCSI uses a Stringprep profile for the names of protocol 231 participants (called initiators and targets). The IQN format of 232 iSCSI names contains a reversed DNS domain name. 233 o SASL and LDAP uses a Stringprep profile for usernames. 234 o LDAP uses a set of Stringprep profiles. 236 The apparent judgement of the BOF attendees [NEWPREP] was that it 237 would be highly desirable to have a replacement of Stringprep, with 238 similar characteristics to IDNA2008. That replacement should be 239 defined so that the protocols could use internationalized strings 240 without a lot of specialized internationalization work, since 241 internationalization expertise is not available in the respective 242 protocols or working groups. Accordingly, the IESG formed the PRECIS 243 working group to undertake the task. 245 Notwithstanding the desire evident in [NEWPREP] and the chartering of 246 a working group, IDNA2008 may be a poor model for what other 247 protocols ought to do, because it is designed to support an old 248 protocol that is designed to operate on the scale of the entire 249 Internet. Moreover, IDNA2008 is intended to be deployed without any 250 change to the base DNS protocol. Other protocols may aim at 251 deployment in more local environments, or may have protocol version 252 negotiation built in. 254 5. Major Topics for Consideration 256 This section provides an overview of major topics that a Stringprep 257 replacement needs to address. The headings correspond roughly with 258 categories under which known Stringprep-using protocol RFCs have been 259 evaluated. For the details of those evaluations, see Appendix A. 261 5.1. Comparison 263 5.1.1. Types of Identifiers 265 Following [I-D.iab-identifier-comparison], it is possible to organize 266 identifiers into three classes in respect of how they may be compared 267 with one another: 269 Absolute Identifiers Identifiers that can be compared byte-by-byte 270 for equality. 271 Definite Identifiers Identifiers that have a well-defined comparison 272 algorithm on which all parties agree. 273 Indefinite Identifiers Identifiers that have no single comparison 274 algorithm on which all parties agree. 276 Definite Identifiers include cases like the comparison of Unicode 277 code points in different encodings: they do not match byte for byte, 278 but can all be converted to a single encoding which then does match 279 byte for byte. Indefinite Identifiers are sometimes algorithmically 280 comparable by well-specified subsets of parties. For more discussion 281 of these categories, see [I-D.iab-identifier-comparison]. 283 The section on treating the existing known cases, Appendix A uses the 284 categories above. 286 5.1.2. Effect of comparison 288 The three classes of comparison style outlined in Section 5.1.1 may 289 have different effects when applied. It is necessary to evaluate the 290 effects if a comparison results in a false positive, and what the 291 effects are if a comparison results in a false negative, especially 292 in terms of the consequences to security and usability. 294 5.2. Dealing with characters 296 This section outlines a range of issues having to do with characters 297 in the target protocols, and outlines the ways in which IDNA2008 298 might be a good analogy to other protocols, and ways in which it 299 might be a poor one. 301 5.2.1. Case folding, case sensitivity, and case preservation 303 In IDNA2003, labels are always mapped to lower case before the 304 Punycode transformation. In IDNA2008, there is no mapping at all: 305 input is either a valid U-label or it is not. At the same time, 306 upper-case characters are by definition not valid U-labels, because 307 they fall into the Unstable category (category B) of [RFC5892]. 309 If there are protocols that require upper and lower cases be 310 preserved, then the analogy with IDNA2008 will break down. 311 Accordingly, existing protocols are to be evaluated according to the 312 following criteria: 314 1. Does the protocol use case folding? For all blocks of code 315 points, or just for certain subsets? 316 2. Is the system or protocol case sensitive? 317 3. Does the system or protocol preserve case? 319 5.2.2. Stringprep and NFKC 321 Stringprep profiles may use normalization. If they do, they use NFKC 322 [UAX15] (most profiles do). It is not clear that NFKC is the right 323 normalization to use in all cases. In [UAX15], there is the 324 following observation regarding Normalization Forms KC and KD: "It is 325 best to think of these Normalization Forms as being like uppercase or 326 lowercase mappings: useful in certain contexts for identifying core 327 meanings, but also performing modifications to the text that may not 328 always be appropriate." In general, it can be said that NFKC is more 329 aggressive about finding matches between codepoints than NFC. For 330 things like the spelling of users' names, then, NFKC may not be the 331 best form to use. At the same time, one of the nice things about 332 NFKC is that it deals with the width of characters that are otherwise 333 similar, by canonicalizing half-width to full-width. This mapping 334 step can be crucial in practice. A replacement for stringprep 335 depends on analyzing the different use profiles and considering 336 whether NFKC or NFC is a better normalization for each profile. 338 For the purposes of evaluating an existing example of Stringprep use, 339 it is helpful to know whether it uses no normalization, NFKC, or NFC. 341 5.2.3. Character mapping 343 Along with the case mapping issues raised in Section 5.2.1, there is 344 the question of whether some characters are mapped either to other 345 characters or to nothing during Stringprep. [RFC3454], Section 3, 346 outlines a number of characters that are mapped to nothing, and also 347 permits Stringprep profiles to define their own mappings. 349 5.2.4. Prohibited characters 351 Along with case folding and other character mappings, many protocols 352 have characters that are simply disallowed. For example, control 353 characters and special characters such as "@" or "/" may be 354 prohibited in a protocol. 356 One of the primary changes of IDNA2008 is in the way it approaches 357 Unicode code points, using the new inclusion-based approach (see 358 Section 1). 360 Because of the default assumption in IDNA2008 that a code point is 361 not allowed by the protocol, it has more than one class of "allowed 362 by the protocol"; this is unlike IDNA2003. While some code points 363 are disallowed outright, some are allowed only in certain contexts. 364 The reasons for the context-dependent rules have to do with the way 365 some characters are used. For instance, the ZERO WIDTH JOINER and 366 ZERO WIDTH NON-JOINER (ZWJ, U+200D and ZWNJ, U+200C) are allowed with 367 contextual rules because they are required in some circumstances, yet 368 are considered punctuation by Unicode and would therefore be 369 DISALLOWED under the usual IDNA2008 derivation rules. The goal of 370 IDNA2008 is to provide the widest repertoire of code points possible 371 and consistent with the traditional DNS "LDH" (letters, digits, 372 hyphen; see [RFC0952]) rule, trusting to the operators of individual 373 zones to make sensible (and usually more restrictive) policies for 374 their zones. 376 5.2.5. Internal structure, delimiters, and special characters 378 IDNA2008 has a special problem with delimiters, because the delimiter 379 "character" in the DNS wire format is not really part of the data. 380 In DNS, labels are not separated exactly; instead, a label carries 381 with it an indicator that says how long the label is. When the label 382 is presented in presentation format as part of a fully qualified 383 domain name, the label separator FULL STOP, U+002E (.) is used to 384 break up the labels. But because that label separator does not 385 travel with the wire format of the domain name, there is no way to 386 encode a different, "internationalized" separator in IDNA2008. 388 Other protocols may include characters with similar special meaning 389 within the protocol. Common characters for these purposes include 390 FULL STOP, U+002E (.); COMMERCIAL AT, U+0040 (@); HYPHEN-MINUS, 391 U+002D (-); SOLIDUS, U+002F (/); and LOW LINE, U+005F (_). The mere 392 inclusion of such a character in the protocol is not enough for it to 393 be considered similar to another protocol using the same character; 394 instead, handling of the character must be taken into consideration 395 as well. 397 An important issue to tackle here is whether it is valuable to map to 398 or from these special characters as part of the Stringprep 399 replacement. In some locales, the analogue to FULL STOP, U+002E is 400 some other character, and users may expect to be able to substitute 401 their normal stop for FULL STOP, U+002E. At the same time, there are 402 predictability arguments in favour of treating identifiers with FULL 403 STOP, U+002E in them just the way they are treated under IDNA2008. 405 5.2.6. Restrictions because of glyph similarity 407 Homoglyphs are similarly (or identically) rendered glyphs of 408 different codepoints. For DNS names, homoglyphs may enable phishing. 409 If a protocol requires some visual comparison by end-users, then the 410 issue of homoglyphs are to be considered. In the DNS context, theses 411 issues are documented in [RFC5894] and [RFC4690]. IDNA2008 does not, 412 however, have a mechanism to deal with them, trusting to DNS zone 413 operators to enact sensible policies for the subset of Unicode they 414 wish to support, given their user community. A similar policy/ 415 protocol split may not be desirable in every protocol. 417 5.3. Where the data comes from and where it goes 419 5.3.1. User input and the source of protocol elements 421 Some protocol elements are provided by users, and others are not. 422 Those that are not may presumably be subject to greater restrictions, 423 whereas those that users provide likely need to permit the broadest 424 range of code points. The following questions are helpful: 426 1. Do users input the strings directly? 427 2. If so, how? (keyboard, stylus, voice, copy-paste, etc.) 428 3. Where do we place the dividing line between user interface and 429 protocol? (see [RFC5895]) 431 5.3.2. User output 433 Just as only some protocol elements are expected to be entered 434 directly by users, only some protocol elements are intended to be 435 consumed directly by users. It is important to know how users are 436 expected to be able to consume the protocol elements, because 437 different environments present different challenges. An element that 438 is only ever delivered as part of a vCard remains in machine-readable 439 format, so the problem of visual confusion is not a great one. Is 440 the protocol element published as part of a vCard, a web directory, 441 on a business card, or on "the side of a bus"? Do users use the 442 protocol element as an identifier (which means that they might enter 443 it again in some other context)? (See also Section 5.2.6.) 445 5.3.3. Operations 447 Some strings are useful as part of the protocol but are not used as 448 input to other operations (for instance, purely informative or 449 descriptive text). Other strings are used directly as input to other 450 operations (such as cryptographic hash functions), or are used 451 together with other strings to (such as concatenating a string with 452 some others to form a unique identifier). 454 5.3.3.1. String classes 456 Strings often have a similar function in different protocols. For 457 instance, many different protocols contain user identifiers or 458 passwords. A single profile for all such uses might be desirable. 460 Often, a string in a protocol is effectively a protocol element from 461 another protocol. For instance, different systems might use the same 462 credentials database for authentication. 464 5.3.3.2. Community Considerations 466 A Stringprep replacement that does anything more than just update 467 Stringprep to the latest version of Unicode will probably entail some 468 changes. It is important to identify the willingness of the 469 protocol-using community to accept backwards-incompatible changes. 470 By the same token, it is important to evaluate the desire of the 471 community for features not available under Stringprep. 473 5.3.3.3. Unicode Incompatible Changes 475 IDNA2008 uses an algorithm to derive the validity of a Unicode code 476 point for use under IDNA2008. It does this by using the properties 477 of each code point to test its validity. 479 This approach depends crucially on the idea that code points, once 480 valid for a protocol profile, will not later be made invalid. That 481 is not a guarantee currently provided by Unicode. Properties of code 482 points may change between versions of Unicode. Rarely, such a change 483 could cause a given code point to become invalid under a protocol 484 profile, even though the code point would be valid with an earlier 485 version of Unicode. This is not merely a theoretical possibility, 486 because it has occurred ([RFC6452]). 488 Accordingly, as in IDNA2008, a Stringprep replacement that intends to 489 be Unicode version agnostic will need to work out a mechanism to 490 address cases where incompatible changes occur because of new Unicode 491 versions. 493 6. Considerations for Stringprep replacement 495 The above suggests the following guidance for replacing Stringprep: 496 o A stringprep replacement should be defined. 497 o The replacement should take an approach similar to IDNA2008, (e.g. 498 by using codepoint properties instead of codepoint whitelisting) 499 in that it enables better Unicode agility. 500 o Protocols share similar characteristics of strings. Therefore, 501 defining internationalization preparation algorithms for the 502 smallest set of string classes may be sufficient for most cases, 503 providing coherence among a set of related protocols or protocols 504 where identifiers are exchanged. 505 o The sets of string classes need to be evaluated according to the 506 considerations that make up the headings in Section 5 507 o It is reasonable to limit scope to Unicode code points, and rule 508 the mapping of data from other character encodings outside the 509 scope of this effort. 510 o The replacement ought at least to provide guidance to applications 511 using the replacement on how to handle protocol incompatibilities 512 resulting from changes to Unicode. In an ideal world, the 513 stringprep replacement would handle the changes automatically, but 514 it appears that such automatic handling would require magic and 515 cannot be expected. 516 o Compatibility within each protocol between a technique that is 517 stringprep-based and the technique's replacement has to be 518 considered very carefully. 520 Existing deployments already depend on Stringprep profiles. 521 Therefore, a replacement must consider the effects of any new 522 strategy on existing deployments. By way of comparison, it is worth 523 noting that some characters were acceptable in IDNA labels under 524 IDNA2003, but are not protocol-valid under IDNA2008 (and conversely); 525 disagreement about what to do during the transition has resulted in 526 different approaches to mapping. Different implementers may make 527 different decisions about what to do in such cases; this could have 528 interoperability effects. It is necessary to trade better support 529 for different linguistic environments against the potential side 530 effects of backward incompatibility. 532 7. Security Considerations 534 This document merely states what problems are to be solved, and does 535 not define a protocol. There are undoubtedly security implications 536 of the particular results that will come from the work to be 537 completed. 539 8. IANA Considerations 541 This document has no actions for IANA. 543 9. Discussion home for this draft 545 Note: RFC-Editor, please remove this section before publication. 547 This document is intended to define the problem space discussed on 548 the precis@ietf.org mailing list. 550 10. Acknowledgements 552 This document is the product of the PRECIS IETF Working Group, and 553 participants in that Working Group were helpful in addressing issues 554 with the text. 556 Specific contributions came from David Black, Alan DeKok, Simon 557 Josefsson, Bill McQuillan, Alexey Melnikov, Peter Saint-Andre, Dave 558 Thaler, and Yoshiro Yoneya. 560 Dave Thaler provided the "buckets" insight in Section 5.1.1, central 561 to the organization of the problem. 563 Evaluations of Stringprep profiles that are included in Appendix B 564 were done by: David Black, Alexey Melnikov, Peter Saint-Andre, Dave 565 Thaler. 567 11. Informative References 569 [I-D.iab-identifier-comparison] 570 Thaler, D., "Issues in Identifier Comparison for Security 571 Purposes", draft-iab-identifier-comparison-04 (work in 572 progress), August 2012. 574 [NEWPREP] "Newprep BoF Meeting Minutes", March 2010. 576 [RFC0952] Harrenstien, K., Stahl, M., and E. Feinler, "DoD Internet 577 host table specification", RFC 952, October 1985. 579 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 580 Requirement Levels", BCP 14, RFC 2119, March 1997. 582 [RFC3454] Hoffman, P. and M. Blanchet, "Preparation of 583 Internationalized Strings ("stringprep")", RFC 3454, 584 December 2002. 586 [RFC3490] Faltstrom, P., Hoffman, P., and A. Costello, 587 "Internationalizing Domain Names in Applications (IDNA)", 588 RFC 3490, March 2003. 590 [RFC3491] Hoffman, P. and M. Blanchet, "Nameprep: A Stringprep 591 Profile for Internationalized Domain Names (IDN)", 592 RFC 3491, March 2003. 594 [RFC3492] Costello, A., "Punycode: A Bootstring encoding of Unicode 595 for Internationalized Domain Names in Applications 596 (IDNA)", RFC 3492, March 2003. 598 [RFC3530] Shepler, S., Callaghan, B., Robinson, D., Thurlow, R., 599 Beame, C., Eisler, M., and D. Noveck, "Network File System 600 (NFS) version 4 Protocol", RFC 3530, April 2003. 602 [RFC3722] Bakke, M., "String Profile for Internet Small Computer 603 Systems Interface (iSCSI) Names", RFC 3722, April 2004. 605 [RFC3748] Aboba, B., Blunk, L., Vollbrecht, J., Carlson, J., and H. 606 Levkowetz, "Extensible Authentication Protocol (EAP)", 607 RFC 3748, June 2004. 609 [RFC3920] Saint-Andre, P., Ed., "Extensible Messaging and Presence 610 Protocol (XMPP): Core", RFC 3920, October 2004. 612 [RFC3922] Saint-Andre, P., "Mapping the Extensible Messaging and 613 Presence Protocol (XMPP) to Common Presence and Instant 614 Messaging (CPIM)", RFC 3922, October 2004. 616 [RFC4011] Waldbusser, S., Saperia, J., and T. Hongal, "Policy Based 617 Management MIB", RFC 4011, March 2005. 619 [RFC4013] Zeilenga, K., "SASLprep: Stringprep Profile for User Names 620 and Passwords", RFC 4013, February 2005. 622 [RFC4279] Eronen, P. and H. Tschofenig, "Pre-Shared Key Ciphersuites 623 for Transport Layer Security (TLS)", RFC 4279, 624 December 2005. 626 [RFC4314] Melnikov, A., "IMAP4 Access Control List (ACL) Extension", 627 RFC 4314, December 2005. 629 [RFC4422] Melnikov, A. and K. Zeilenga, "Simple Authentication and 630 Security Layer (SASL)", RFC 4422, June 2006. 632 [RFC4505] Zeilenga, K., "Anonymous Simple Authentication and 633 Security Layer (SASL) Mechanism", RFC 4505, June 2006. 635 [RFC4511] Sermersheim, J., "Lightweight Directory Access Protocol 636 (LDAP): The Protocol", RFC 4511, June 2006. 638 [RFC4513] Harrison, R., "Lightweight Directory Access Protocol 639 (LDAP): Authentication Methods and Security Mechanisms", 640 RFC 4513, June 2006. 642 [RFC4518] Zeilenga, K., "Lightweight Directory Access Protocol 643 (LDAP): Internationalized String Preparation", RFC 4518, 644 June 2006. 646 [RFC4616] Zeilenga, K., "The PLAIN Simple Authentication and 647 Security Layer (SASL) Mechanism", RFC 4616, August 2006. 649 [RFC4643] Vinocur, J. and K. Murchison, "Network News Transfer 650 Protocol (NNTP) Extension for Authentication", RFC 4643, 651 October 2006. 653 [RFC4683] Park, J., Lee, J., Lee, H., Park, S., and T. Polk, 654 "Internet X.509 Public Key Infrastructure Subject 655 Identification Method (SIM)", RFC 4683, October 2006. 657 [RFC4690] Klensin, J., Faltstrom, P., Karp, C., and IAB, "Review and 658 Recommendations for Internationalized Domain Names 659 (IDNs)", RFC 4690, September 2006. 661 [RFC4790] Newman, C., Duerst, M., and A. Gulbrandsen, "Internet 662 Application Protocol Collation Registry", RFC 4790, 663 March 2007. 665 [RFC4954] Siemborski, R. and A. Melnikov, "SMTP Service Extension 666 for Authentication", RFC 4954, July 2007. 668 [RFC5034] Siemborski, R. and A. Menon-Sen, "The Post Office Protocol 669 (POP3) Simple Authentication and Security Layer (SASL) 670 Authentication Mechanism", RFC 5034, July 2007. 672 [RFC5051] Crispin, M., "i;unicode-casemap - Simple Unicode Collation 673 Algorithm", RFC 5051, October 2007. 675 [RFC5054] Taylor, D., Wu, T., Mavrogiannopoulos, N., and T. Perrin, 676 "Using the Secure Remote Password (SRP) Protocol for TLS 677 Authentication", RFC 5054, November 2007. 679 [RFC5122] Saint-Andre, P., "Internationalized Resource Identifiers 680 (IRIs) and Uniform Resource Identifiers (URIs) for the 681 Extensible Messaging and Presence Protocol (XMPP)", 682 RFC 5122, February 2008. 684 [RFC5280] Cooper, D., Santesson, S., Farrell, S., Boeyen, S., 685 Housley, R., and W. Polk, "Internet X.509 Public Key 686 Infrastructure Certificate and Certificate Revocation List 687 (CRL) Profile", RFC 5280, May 2008. 689 [RFC5456] Spencer, M., Capouch, B., Guy, E., Miller, F., and K. 690 Shumard, "IAX: Inter-Asterisk eXchange Version 2", 691 RFC 5456, February 2010. 693 [RFC5661] Shepler, S., Eisler, M., and D. Noveck, "Network File 694 System (NFS) Version 4 Minor Version 1 Protocol", 695 RFC 5661, January 2010. 697 [RFC5802] Newman, C., Menon-Sen, A., Melnikov, A., and N. Williams, 698 "Salted Challenge Response Authentication Mechanism 699 (SCRAM) SASL and GSS-API Mechanisms", RFC 5802, July 2010. 701 [RFC5804] Melnikov, A. and T. Martin, "A Protocol for Remotely 702 Managing Sieve Scripts", RFC 5804, July 2010. 704 [RFC5890] Klensin, J., "Internationalized Domain Names for 705 Applications (IDNA): Definitions and Document Framework", 706 RFC 5890, August 2010. 708 [RFC5891] Klensin, J., "Internationalized Domain Names in 709 Applications (IDNA): Protocol", RFC 5891, August 2010. 711 [RFC5892] Faltstrom, P., "The Unicode Code Points and 712 Internationalized Domain Names for Applications (IDNA)", 713 RFC 5892, August 2010. 715 [RFC5893] Alvestrand, H. and C. Karp, "Right-to-Left Scripts for 716 Internationalized Domain Names for Applications (IDNA)", 717 RFC 5893, August 2010. 719 [RFC5894] Klensin, J., "Internationalized Domain Names for 720 Applications (IDNA): Background, Explanation, and 721 Rationale", RFC 5894, August 2010. 723 [RFC5895] Resnick, P. and P. Hoffman, "Mapping Characters for 724 Internationalized Domain Names in Applications (IDNA) 725 2008", RFC 5895, September 2010. 727 [RFC6452] Faltstrom, P. and P. Hoffman, "The Unicode Code Points and 728 Internationalized Domain Names for Applications (IDNA) - 729 Unicode 6.0", RFC 6452, November 2011. 731 [UAX15] "Unicode Standard Annex #15: Unicode Normalization Forms", 732 UAX 15, September 2009. 734 [Unicode61] 735 The Unicode Consortium. The Unicode Standard, Version 736 6.1, defined by:, "The Unicode Standard -- Version 6.1", 737 (Mountain View, CA: The Unicode Consortium, 2012. ISBN 738 978-1-936213-02-3), September 2009, 739 . 741 [ietf78precis] 742 Blanchet, M., "PRECIS Framework", Proceedings of the 743 Seventy-Eighth Internet Engineering Task 744 Force https://www.ietf.org/proceedings/78/, July 2010, 745 . 747 Appendix A. Classification of Stringprep Profiles 749 A number of the known cases of Stringprep use were evaluated during 750 the preparation of this document. The known cases are here described 751 in two ways. The types of identifiers the protocol uses is first 752 called out in the ID type column (from Section 5.1.1), using the 753 short forms "a" for Absolute, "d" for Definite, and "i" for 754 Indefinite. Next, there is a column that contains an "i" if the 755 protocol string comes from user input, an "o" if the protocol string 756 becomes user-facing output, "b" if both are true, and "n" if neither 757 is true. 759 +------+--------+-------+ 760 | RFC | IDtype | User? | 761 +------+--------+-------+ 762 | 3722 | a | b | 763 | 3748 | - | - | 764 | 3920 | a,d | b | 765 | 4505 | a | i | 766 | 4314 | a,d | b | 767 | 4954 | a,d | b | 768 | 5034 | a,d | b | 769 | 5804 | a,d | b | 770 +------+--------+-------+ 772 Table 1 774 Appendix B. Evaluation of Stringprep Profiles 776 This section is a summary of evaluation of Stringprep profiles that 777 was done to get a good understanding of the usage of Stringprep. 778 This summary is by no means normative nor the actual evaluations 779 themselves. A template was used for reviewers to get a coherent view 780 of all evaluations. 782 B.1. iSCSI Stringprep Profile: RFC3722 (and RFC3721, RFC3720) 784 Description: An iSCSI session consists of an initiator (i.e., host 785 or server that uses storage) communicating with a target (i.e., a 786 storage array or other system that provides storage). Both the 787 iSCSI initiator and target are named by iSCSI Names. The iSCSI 788 stringprep profile is used for iSCSI names. 789 How it is used: iSCSI initiators and targets (see above). They can 790 also be used to identify SCSI ports (these are software entities 791 in the iSCSI protocol, not hardware ports), and iSCSI logical 792 units (storage volumes), although both are unusual in practice. 793 What entities create these identifiers? Generally a Human user (1) 794 configures an Automated system (2) that generates the names. 795 Advance configuration of the system is required due to the 796 embedded use of external unique identifier (from the DNS or IEEE). 797 How is the string input in the system? Keyboard and copy-paste are 798 common. Copy-paste is common because iSCSI names are long enough 799 to be problematic for humans to remember, causing use of email, 800 sneaker-net, text files, etc. to avoid mistype mistakes. 801 Where do we place the dividing line between user interface and 802 protocol? The iSCSI protocol requires that all internationalization 803 string preparation occur in the user interface. The iSCSI 804 protocol treats iSCSI names as opaque identifiers that are 805 compared byte-by-byte for equality. iSCSI names are generally not 806 checked for correct formatting by the protocol. 807 What entities enforce the rules? There are no iSCSI-specific 808 enforcement entities, although the use of unique identifier 809 information in the names relies on DNS registrars and the IEEE 810 Registration Authority. 811 Comparison Byte-by-byte 812 Case Folding, Sensitivity, Preservation Case folding is required for 813 the code blocks specified in RFC 3454, Table B.2. The overall 814 iSCSI naming system (UI + protocol) is case-insensitive. 815 What is the impact if the comparison results in a false positive? 816 Potential access to the wrong storage. - If the initiator has no 817 access to the wrong storage, an authentication failure is the 818 probable result. - If the initiator has access to the wrong 819 storage, the resulting mis-identification could result in use of 820 the wrong data and possible corruption of stored data. 821 What is the impact if the comparison results in a false negative? 822 Denial of authorized storage access. 823 What are the security impacts? iSCSI names may be used as the 824 authentication identities for storage systems. Comparison 825 problems could result in authentication problems, although note 826 that authentication failure ameliorates some of the false positive 827 cases. 828 Normalization NFKC, as specified by RFC 3454. 829 Mapping Yes, as specified by table B.1 in RFC 3454 830 Disallowed Characters Only the following characters are allowed: - 831 ASCII dash, dot, colon - ASCII lower case letters and digits - 832 Unicode lower case characters as specified by RFC 3454 All other 833 characters are disallowed. 834 Which other strings or identifiers are these most similar to? None - 835 iSCSI names are unique to iSCSI. 836 Are these strings or identifiers sometimes the same as strings or 837 identifiers from other protocols? No 838 Does the identifier have internal structure that needs to be 839 respected? Yes - ASCII dot, dash and colon are used for internal 840 name structure. These are not reserved characters in that they 841 can occur in the name in locations other than those used for 842 structuring purposes (e.g., only the first occurrence of a colon 843 character is structural, others are not). 844 How are users exposed to these strings? How are they published? 845 iSCSI names appear in server and storage system configuration 846 interfaces. They also appear in system logs. 847 Is the string / identifier used as input to other operations? 848 Effectively, no. The rarely used port and logical unit names 849 involve concatenation, which effectively extends a unique iSCSI 850 Name for a target to uniquely identify something within that 851 target. 853 How much tolerance for change from existing stringprep approach? 854 Good tolerance; the community would prefer that 855 internationalization experts solve internationalization problems. 856 How strong a desire for change (e.g., for Unicode agility)? Unicode 857 agility is desired in principle as long as nothing significant 858 breaks. 860 B.2. SMTP/POP3/ManageSieve Stringprep Profiles: RFC4954,RFC5034,RFC 861 5804 863 Description: Authorization identity (user identifier) exchanged 864 during SASL authentication: AUTH (SMTP/POP3) or AUTHENTICATE 865 (ManageSieve) command. 866 How It's Used: Used for proxy authorization, e.g. to [lawfully] 867 impersonate a particular user after a privileged authentication 868 Who Generates It: Typically generated by email system administrators 869 using some tools/conventions, sometimes from some backend 870 database. - In some setups human users can register own usernames 871 (e.g. webmail self registration) 872 User Input Methods: - Typed by user / selected from a list - Copy- 873 and-paste - Perhaps voice input - Can also be specified in 874 configuration files or on a command line 875 Enforcement: - Rules enforced by server / add-on service (e.g., 876 gateway service) on registration of account 877 Comparison Method: "Type 1" (byte-for-byte) or "type 2" (compare by 878 a common algorithm that everyone agrees on (e.g., normalize and 879 then compare the result byte-by-byte)) 880 Case Folding, Sensitivity, Preservation: Most likely case sensitive. 881 Exact requirements on case-sensitivity/case-preservation depend on 882 a specific implementation, e.g. an implementation might treat all 883 user identifiers as case insensitive (or case insensitive for US- 884 ASCII subset only). 885 Impact of Comparison: False positives: - an unauthorized user is 886 allowed email service access (login) False negatives: - an 887 authorized user is denied email service access 888 Normalization: NFKC (as per RFC 4013) 889 Mapping: (see Section 2 of RFC 4013 for the full list): Non ASCII 890 spaces are mapped to space, etc. 891 Disallowed Characters: (see Section 2 of RFC 4013 for the full 892 list): Unicode Control characters, etc. 893 String Classes: - simple username. See Section 2 of RFC 4013 for 894 details on restrictions. Note that some implementations allow 895 spaces in these. While implementations are not required to use a 896 specific format, an authorization identity frequently has the same 897 format as an email address (and EAI email address in the future), 898 or as a left hand side of an email address. Note: whatever is 899 recommended for SMTP/POP/ManageSieve authorization identity should 900 also be used for IMAP authorization identities, as IMAP/POP3/SMTP/ 901 ManageSieve are frequently implemented together. 902 Internal Structure: None 903 User Output: Unlikely, but possible. For example, if it is the same 904 as an email address. 905 Operations: - Sometimes concatenated with other data and then used 906 as input to a cryptographic hash function 907 How much tolerance for change from existing stringprep approach? Not 908 sure. 909 Background information: In RFC 5034, when describing the POP3 AUTH 910 command: The authorization identity generated by the SASL exchange 911 is a simple username, and SHOULD use the SASLprep profile (see 912 RFC4013) of the StringPrep algorithm (see RFC3454) to prepare 913 these names for matching. If preparation of the authorization 914 identity fails or results in an empty string (unless it was 915 transmitted as the empty string), the server MUST fail the 916 authentication. In RFC 4954, when describing the SMTP AUTH 917 command: The authorization identity generated by this SASL 918 exchange is a "simple username" (in the sense defined in 919 SASLprep), and both client and server SHOULD (*) use the SASLprep 920 profile of the StringPrep algorithm to prepare these names for 921 transmission or comparison. If preparation of the authorization 922 identity fails or results in an empty string (unless it was 923 transmitted as the empty string), the server MUST fail the 924 authentication. (*) Note: Future revision of this specification 925 may change this requirement to MUST. Currently, the SHOULD is 926 used in order to avoid breaking the majority of existing 927 implementations. In RFC 5804, when describing the ManageSieve 928 AUTHENTICATE command: The authorization identity generated by this 929 SASL exchange is a "simple username" (in the sense defined in 930 SASLprep), and both client and server MUST use the SASLprep 931 profile of the StringPrep algorithm to prepare these names for 932 transmission or comparison. If preparation of the authorization 933 identity fails or results in an empty string (unless it was 934 transmitted as the empty string), the server MUST fail the 935 authentication. 937 B.3. IMAP Stringprep Profiles: RFC5738, RFC4314: Usernames 939 Evaluation Note These documents have 2 types of strings (usernames 940 and passwords), so there are two separate templates. 941 Description: "username" parameter to the IMAP LOGIN command, 942 identifiers in IMAP ACL commands. Note that any valid username is 943 also an IMAP ACL identifier, but IMAP ACL identifiers can include 944 other things like name of group of users. 946 How It's Used: Used for authentication (Usernames), or in IMAP 947 Access Control Lists (Usernames or Group names) 948 Who Generates It: - Typically generated by email system 949 administrators using some tools/conventions, sometimes from some 950 backend database. - In some setups human users can register own 951 usernames (e.g. webmail self registration) 952 User Input Methods: - Typed by user / selected from a list - Copy- 953 and-paste - Perhaps voice input - Can also be specified in 954 configuration files or on a command line 955 Enforcement: - Rules enforced by server / add-on service (e.g., 956 gateway service) on registration of account 957 Comparison Method: Type 1" (byte-for-byte) or "type 2" (compare by a 958 common algorithm that everyone agrees on (e.g., normalize and then 959 compare the result byte-by-byte)) 960 Case Folding, Sensitivity, Preservation: - Most likely case 961 sensitive. Exact requirements on case-sensitivity/ 962 case-preservation depend on a specific implementation, e.g. an 963 implementation might treat all user identifiers as case 964 insensitive (or case insensitive for US-ASCII subset only). 965 Impact of Comparison: False positives: - an unauthorized user is 966 allowed IMAP access (login) - improperly grant privileges (e.g., 967 access to a specific mailbox, ability to manage ACLs for a 968 mailbox) False negatives: - an authorized user is denied IMAP 969 access - unable to use granted privileges (e.g., access to a 970 specific mailbox, ability to manage ACLs for a mailbox) 971 Normalization: NFKC (as per RFC 4013) 972 Mapping: (see Section 2 of RFC 4013 for the full list): non ASCII 973 spaces are mapped to space 974 Disallowed Characters: (see Section 2 of RFC 4013 for the full 975 list): Unicode Control characters, etc. 976 String Classes: - simple username. See Section 2 of RFC 4013 for 977 details on restrictions. Note that some implementations allow 978 spaces in these. While IMAP implementations are not required to 979 use a specific format, an IMAP username frequently has the same 980 format as an email address (and EAI email address in the future), 981 or as a left hand side of an email address. Note: whatever is 982 recommended for IMAP username should also be used for ManageSieve, 983 POP3 and SMTP authorization identities, as IMAP/POP3/SMTP/ 984 ManageSieve are frequently implemented together. 985 Internal Structure: None 986 User Output: Unlikely, but possible. For example, if it is the same 987 as an email address. - access control lists (e.g. in IMAP ACL 988 extension), both when managing membership and listing membership 989 of existing access control lists. - often show up as mailbox names 990 (under Other Users IMAP namespace) 992 Operations: - Sometimes concatenated with other data and then used 993 as input to a cryptographic hash function 994 How much tolerance for change from existing stringprep approach? Not 995 sure. Non-ASCII IMAP usernames are currently prohibited by IMAP 996 (RFC 3501). However they are allowed when used in IMAP ACL 997 extension. 999 B.4. IMAP Stringprep Profiles: RFC5738: Passwords 1001 Description: "Password" parameter to the IMAP LOGIN command 1002 How It's Used: Used for authentication (Passwords) 1003 Who Generates It: Either generated by email system administrators 1004 using some tools/conventions, or specified by the human user. 1005 User Input Methods: - Typed by user - Copy-and-paste - Perhaps voice 1006 input - Can also be specified in configuration files or on a 1007 command line 1008 Enforcement: Rules enforced by server / add-on service (e.g., 1009 gateway service or backend databse) on registration of account 1010 Comparison Method: "Type 1" (byte-for-byte) 1011 Case Folding, Sensitivity, Preservation: Most likely case sensitive. 1012 Impact of Comparison: False positives: - an unauthorized user is 1013 allowed IMAP access (login) False negatives: - an authorized user 1014 is denied IMAP access 1015 Normalization: NFKC (as per RFC 4013) 1016 Mapping: (see Section 2 of RFC 4013 for the full list): non ASCII 1017 spaces are mapped to space 1018 Disallowed Characters: (see Section 2 of RFC 4013 for the full 1019 list): Unicode Control characters, etc. 1020 String Classes: Currently defined as "simple username" (see Section 1021 2 of RFC 4013 for details on restrictions.), however this is 1022 likely to be a different class from usernames. Note that some 1023 implementations allow spaces in these. Password in all email 1024 related protocols should be treated in the same way. Same 1025 passwords are frequently shared with web, IM, etc. applications. 1026 Internal Structure: None 1027 User Output: - text of email messages (e.g. in "you forgot your 1028 password" email messages) - web page / directory - side of the bus 1029 / in ads -- possible 1030 Operations: Sometimes concatenated with other data and then used as 1031 input to a cryptographic hash function. Frequently stored as is, 1032 or hashed. 1033 How much tolerance for change from existing stringprep approach? Not 1034 sure. Non-ASCII IMAP passwords are currently prohibited by IMAP 1035 (RFC 3501), however they are likely to be in widespread use. 1037 Background information: RFC 5738 (IMAP INTERNATIONALIZATION): 5. 1038 UTF8=USER Capability If the "UTF8=USER" capability is advertised, 1039 that indicates the server accepts UTF-8 user names and passwords 1040 and applies SASLprep RFC4013 to both arguments of the LOGIN 1041 command. The server MUST reject UTF-8 that fails to comply with 1042 the formal syntax in RFC 3629 RFC3629 or if it encounters Unicode 1043 characters listed in Section 2.3 of SASLprep RFC 4013 RFC4013. 1044 RFC 4314 (IMAP4 Access Control List (ACL) Extension): 3. Access 1045 control management commands and responses Servers, when processing 1046 a command that has an identifier as a parameter (i.e., any of 1047 SETACL, DELETEACL, and LISTRIGHTS commands), SHOULD first prepare 1048 the received identifier using "SASLprep" profile SASLprep of the 1049 "stringprep" algorithm Stringprep. If the preparation of the 1050 identifier fails or results in an empty string, the server MUST 1051 refuse to perform the command with a BAD response. Note that 1052 Section 6 recommends additional identifier's verification steps. 1053 and in Section 6: This document relies on SASLprep to describe 1054 steps required to perform identifier canonicalization 1055 (preparation). The preparation algorithm in SASLprep was 1056 specifically designed such that its output is canonical, and it is 1057 well-formed. However, due to an anomaly PR29 in the specification 1058 of Unicode normalization, canonical equivalence is not guaranteed 1059 for a select few character sequences. Identifiers prepared with 1060 SASLprep can be stored and returned by an ACL server. The anomaly 1061 affects ACL manipulation and evaluation of identifiers containing 1062 the selected character sequences. These sequences, however, do 1063 not appear in well-formed text. In order to address this problem, 1064 an ACL server MAY reject identifiers containing sequences 1065 described in PR29 by sending the tagged BAD response. This is in 1066 addition to the requirement to reject identifiers that fail 1067 SASLprep preparation as described in Section 3. 1069 B.5. Anonymous SASL Stringprep Profiles: RFC4505 1071 Description: RFC 4505 defines a "trace" field: 1072 Comparison: this field is not intended for comparison (only used for 1073 logging) 1074 Case folding; case sensitivity, preserve case: No case folding/case 1075 sensitive 1076 Do users input the strings directly? Yes. Possibly entered in 1077 configuration UIs, or on a command line. Can also be stored in 1078 configuration files. The value can also be automatically 1079 generated by clients (e.g. a fixed string is used, or a user's 1080 email address). 1082 How users input strings? Keyboard/voice, stylus (pick from a list). 1083 Copy-paste - possibly. 1084 Normalization: None 1085 Disallowed Characters Control characters are disallowed. (See 1086 Section 3 of RFC 4505) 1087 Which other strings or identifiers are these most similar to? RFC 1088 4505 says that the trace "should take one of two forms: an 1089 Internet email address, or an opaque string that does not contain 1090 the '@' U+0040) character and that can be interpreted by the 1091 system administrator of the client's domain." In practice, this 1092 is a freeform text, so it belongs to a different class from "email 1093 address" or "username". 1094 Are these strings or identifiers sometimes the same as strings or 1095 identifiers from other protocols (e.g., does an IM system sometimes 1096 use the same credentials database for authentication as an email 1097 system)? Yes: see above. However there is no strong need to keep 1098 them consistent in the future. 1099 How are users exposed to these strings, how are they published? No. 1100 However, The value can be seen in server logs 1101 Impacts of false positives and false negatives: False positive: a 1102 user can be confused with another user. False negative: two 1103 distinct users are treated as the same user. But note that the 1104 trace field is not authenticated, so it can be easily falsified. 1105 Tolerance of changes in the community The community would be 1106 flexible. 1107 Delimiters No internal structure, but see comments above about 1108 frequent use of email addresses. 1109 Background information: The Anonymous Mechanism The mechanism 1110 consists of a single message from the client to the server. The 1111 client may include in this message trace information in the form 1112 of a string of UTF-8-encoded Unicode characters prepared in 1113 accordance with StringPrep and the "trace" stringprep profile 1114 defined in Section 3 of this document. The trace information, 1115 which has no semantical value, should take one of two forms: an 1116 Internet email address, or an opaque string that does not contain 1117 the '@' (U+0040) character and that can be interpreted by the 1118 system administrator of the client's domain. For privacy reasons, 1119 an Internet email address or other information identifying the 1120 user should only be used with permission from the user. 3. The 1121 "trace" Profile of "Stringprep" This section defines the "trace" 1122 profile of StringPrep. This profile is designed for use with the 1123 SASL ANONYMOUS Mechanism. Specifically, the client is to prepare 1124 the message production in accordance with this profile. The 1125 character repertoire of this profile is Unicode 3.2. No mapping 1126 is required by this profile. No Unicode normalization is required 1127 by this profile. The list of unassigned code points for this 1128 profile is that provided in Appendix A of StringPrep. Unassigned 1129 code points are not prohibited. Characters from the following 1130 tables of StringPrep are prohibited: - C.2.1 (ASCII control 1131 characters) - C.2.2 (Non-ASCII control characters) - C.3 (Private 1132 use characters) - C.4 (Non-character code points) - C.5 (Surrogate 1133 codes) - C.6 (Inappropriate for plain text) - C.8 (Change display 1134 properties are deprecated) - C.9 (Tagging characters) No 1135 additional characters are prohibited. This profile requires 1136 bidirectional character checking per Section 6 of StringPrep. 1138 B.6. XMPP Stringprep Profiles: RFC3920 Nodeprep 1140 Description: Localpart of JabberID ("JID"), as in: 1141 localpart@domainpart/resourcepart 1142 How It's Used: - Usernames (e.g., stpeter@jabber.org) - Chatroom 1143 names (e.g., precis@jabber.ietf.org) - Publish-subscribe nodes - 1144 Bot names 1145 Who Generates It: - Typically, end users via an XMPP client - 1146 Sometimes created in an automated fashion 1147 User Input Methods: - Typed by user - Copy-and-paste - Perhaps voice 1148 input - Clicking a URI/IRI 1149 Enforcement: - Rules enforced by server / add-on service (e.g., 1150 chatroom service) on registration of account, creation of room, 1151 etc. 1152 Comparison Method: "Type 2" (common algorithm) 1153 Case Folding, Sensitivity, Preservation: - Strings are always folded 1154 to lowercase - Case is not preserved 1155 Impact of Comparison: False positives: - unable to authenticate at 1156 server (or authenticate to wrong account) - add wrong person to 1157 buddy list - join the wrong chatroom - improperly grant privileges 1158 (e.g., chatroom admin) - subscribe to wrong pubsub node - interact 1159 with wrong bot - allow communication with blocked entity False 1160 negatives: - unable to authenticate - unable to add someone to 1161 buddy list - unable to join desired chatroom - unable to use 1162 granted privileges (e.g., chatroom admin) - unable to subscribe to 1163 desired pubsub node - unable to interact with desired bot - 1164 disallow communication with unblocked entity 1165 Normalization: NFKC 1166 Mapping: Spaces are mapped to nothing 1167 Disallowed Characters: ",&,',/,:,<,>,@ 1168 String Classes: - Often similar to generic username - Often similar 1169 to localpart of email address - Sometimes same as localpart of 1170 email address 1171 Internal Structure: None 1172 User Output: - vCard - email signature - web page / directory - text 1173 of message (e.g., in a chatroom) 1175 Operations: - Sometimes concatenated with other data and then used 1176 as input to a cryptographic hash function 1178 B.7. XMPP Stringprep Profiles: RFC3920 Resourceprep 1180 Description: - Resourcepart of JabberID ("JID"), as in: 1181 localpart@domainpart/resourcepart - Typically free-form text 1182 How It's Used: - Device / session names (e.g., 1183 stpeter@jabber.org/Home) - Nicknames (e.g., 1184 precis@jabber.ietf.org/StPeter) 1185 Who Generates It: - Often human users via an XMPP client - Often 1186 generated in an automated fashion by client or server 1187 User Input Methods: - Typed by user - Copy-and-paste - Perhaps voice 1188 input - Clicking a URI/IRI 1189 Enforcement: - Rules enforced by server / add-on service (e.g., 1190 chatroom service) on account login, joining a chatroom, etc. 1191 Comparison Method: "Type 2" (byte-for-byte) 1192 Case Folding, Sensitivity, Preservation: - Strings are never folded 1193 - Case is preserved 1194 Impact of Comparison: False positives: - interact with wrong device 1195 (e.g., for file transfer or voice call) - interact with wrong 1196 chatroom participant - improperly grant privileges (e.g., chatroom 1197 moderator) - allow communication with blocked entity False 1198 negatives: - unable to choose desired chatroom nick - unable to 1199 use granted privileges (e.g., chatroom moderator) - disallow 1200 communication with unblocked entity 1201 Normalization: NFKC 1202 Mapping: Spaces are mapped to nothing 1203 Disallowed Characters: None 1204 String Classes: Basically a free-form identifier 1205 Internal Structure: None 1206 User Output: - text of message (e.g., in a chatroom) - device names 1207 often not exposed to human users 1208 Operations: Sometimes concatenated with other data and then used as 1209 input to a cryptographic hash function 1211 B.8. EAP Stringprep Profiles: RFC3748 1213 Description: RFC 3748 section 5 references Stringprep, but the WG 1214 did not agree with the text (was added by IESG) and there are no 1215 known implementations that use Stringprep. The main problem with 1216 that text is that the use of strings is a per-method concept, not 1217 a generic EAP concept and so RFC 3748 itself does not really use 1218 Stringprep, but individual EAP methods could. As such, the 1219 answers to the template questions are mostly not applicable, but a 1220 few answers are universal across methods. The list of IANA 1221 registered EAP methods is at http://www.iana.org/assignments/ 1222 eap-numbers/eap-numbers.xml#eap-numbers-3 1224 Comparison Methods: n/a (per-method) 1225 Case Folding, Case Sensitivity, Case Preservation: n/a (per-method) 1226 Impact of comparison: A false positive results in unauthorized 1227 network access (and possibly theft of service if some else is 1228 billed). A false negative results in lack of authorized network 1229 access (no connectivity). 1230 User input: n/a (per-method) 1231 Normalization: n/a (per-method) 1232 Mapping: n/a (per-method) 1233 Disallowed characters: n/a (per-method) 1234 String classes: Although some EAP methods may use a syntax similar 1235 to other types of identifiers, EAP mandates that the actual values 1236 must not be assumed to be identifiers usable with anything else. 1237 Internal structure: n/a (per-method) 1238 User output: Identifiers are never human displayed except perhaps as 1239 they're typed by a human. 1240 Operations: n/a (per-method) 1241 Community considerations: There is no resistance to change for the 1242 base EAP protocol (as noted, the WG didn't want the existing 1243 text). However actual use of stringprep, if any, within specific 1244 EAP methods may have resistance. It is currently unknown whether 1245 any EAP methods use stringprep. 1247 Appendix C. Changes between versions 1249 Note to RFC Editor: This section should be removed prior to 1250 publication. 1252 C.1. 00 1254 First WG version. Based on 1255 draft-blanchet-precis-problem-statement-00. 1257 C.2. 01 1259 o Made clear that the document is talking only about Unicode code 1260 points, and not any particular encoding. 1261 o Substantially reorganized the document along the lines of the 1262 review template at . 1264 o Included specific questions for each topic for consideration. 1265 o Moved spot for individual protocol review to appendix. Not 1266 populated yet. 1268 C.3. 02 1270 o Cleared up details of comparison classes 1271 o Added a section on changes in Unicode 1273 C.4. 03 1275 o Aligned comparison discussion with identifier discussion from 1276 draft-iab-identifier-comparison-00 1277 o Added section on classes of strings ("Namey" and so on) 1279 C.5. 04 1281 Keepalive version 1283 C.6. 05 1285 o Changed classes of strings to align with framework doc 1286 o Altered table in Appendix A 1287 o Added all profiles evaluations from the wg wiki in appendix B 1289 C.7. 06 1291 o Respond to comments received in WGLC 1292 o Removed classes of strings (also from Appendix A) 1293 o Moved inclusion/exclusion distinction to Introduction 1294 o Fix some sentences to clarify terminology and add or fix 1295 references 1297 Authors' Addresses 1299 Marc Blanchet 1300 Viagenie 1301 246 Aberdeen 1302 Quebec, QC G1R 2E1 1303 Canada 1305 Email: Marc.Blanchet@viagenie.ca 1306 URI: http://viagenie.ca 1307 Andrew Sullivan 1308 Dyn, Inc. 1309 150 Dow St 1310 Manchester, NH 03101 1311 U.S.A. 1313 Email: asullivan@dyn.com